1 Adam Smith’s Theory of the Persistence of Slavery And its Abolition in Western Europe Barry R. Weingast 1 Stanford University July 2015 Abstract Adam Smith made two positive claims about slavery in the context of developing economies. First, Smith explains that slavery is in general highly inefficient. By his account, the net product under freedom is 12 times larger than under slavery. Second, he observes that, despite its inefficiencies, slavery persists in most of the world. Taken together, these claims create a fundamental puzzle: Why do elites – owning slaves and holding political control – fail to make themselves better off by freeing their slaves? Smith gives two very different answers to this puzzle. The first is psychological. Smith asserts that people have a fundamental desire to dominate others, and slavery provided that opportunity for slaveholding elites. The first explanation is the most commonly advanced in the literature. Yet no where else does Smith use the assumption of domination. This explanation therefore seems ad hoc. I favor instead Smith’s second explanation. This argument, far less known, involves commitment problems. Freeing the slaves would deprive slaveholders of their property. How would they be compensated? In principle, a long-term compensation scheme could solve this problem. But in the undeveloped societies Smith discusses, such as feudal Europe, long-term contracts were difficult to enforce. Indeed, I show that both parties to the long-term compensation scheme had incentives to dishonor it. In the presence of commitment problems, masters could not be assured they would, in fact, be better off freeing their slaves. Slaveholders therefore rationally avoided emancipation despite its inefficiency 1. Introduction Smith forcefully argued that slavery was highly inefficient, implying that freeing the slaves was Pareto improving: both slaves and their masters could be made better off without slavery. Yet Smith also observed that slavery was abolished only in Western Europe, a small “corner” of the world. 1 Senior Fellow, Hoover Institution; and Ward C. Krebs Family Professor, Department of Political Science, Stanford University. The author gratefully acknowledges Glory Liu and Josiah Ober for helpful conversations.
31
Embed
Adam Smith’s Theory of the Persistence of Slavery And its ... · PDF file1 Adam Smith’s Theory of the Persistence of Slavery And its Abolition in Western Europe Barry R. Weingast1
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
Adam Smith’s Theory of the Persistence of Slavery And its Abolition in Western Europe
Barry R. Weingast1 Stanford University
July 2015
Abstract
Adam Smith made two positive claims about slavery in the context of developing economies. First, Smith
explains that slavery is in general highly inefficient. By his account, the net product under freedom is 12
times larger than under slavery. Second, he observes that, despite its inefficiencies, slavery persists in
most of the world. Taken together, these claims create a fundamental puzzle: Why do elites – owning
slaves and holding political control – fail to make themselves better off by freeing their slaves?
Smith gives two very different answers to this puzzle. The first is psychological. Smith asserts that people
have a fundamental desire to dominate others, and slavery provided that opportunity for slaveholding
elites. The first explanation is the most commonly advanced in the literature. Yet no where else does
Smith use the assumption of domination. This explanation therefore seems ad hoc.
I favor instead Smith’s second explanation. This argument, far less known, involves commitment
problems. Freeing the slaves would deprive slaveholders of their property. How would they be
compensated? In principle, a long-term compensation scheme could solve this problem. But in the
undeveloped societies Smith discusses, such as feudal Europe, long-term contracts were difficult to
enforce. Indeed, I show that both parties to the long-term compensation scheme had incentives to
dishonor it. In the presence of commitment problems, masters could not be assured they would, in fact, be
better off freeing their slaves. Slaveholders therefore rationally avoided emancipation despite its
inefficiency
1. Introduction
Smith forcefully argued that slavery was highly inefficient, implying that freeing the
slaves was Pareto improving: both slaves and their masters could be made better off without
slavery. Yet Smith also observed that slavery was abolished only in Western Europe, a small
“corner” of the world.
1 Senior Fellow, Hoover Institution; and Ward C. Krebs Family Professor, Department of Political Science,
Stanford University. The author gratefully acknowledges Glory Liu and Josiah Ober for helpful conversations.
2
Smith’s view presents two puzzles. First, if slavery was so inefficient, why did it persist
around the world and in Western Europe through the feudal period? Second, how was slavery
abolished in Western Europe?
The purpose of this paper is to use modern tools from economics and political science to
understand Smith’s approach to slavery and his answers to the puzzles. In brief, Smith provides
two answers to the first puzzle. He first offers a psychological explanation, arguing that people,
masters in particular, have a “love of domination” (LJ(A) 2
114:186). Even though slavery is
economically inefficient, masters in effect indulge their preferences for domination. Most
scholars in the modern literature take this perspective as Smith's explanation (see, for example,
Brown 2010, Griswold 1999, 199-201; Pack 1991,130-31).3
Smith’s second argument, which occurs on the page following the first argument (LJ(A)
116:187), involves difficulties with compensating the lords who owned the slaves. Abolition
would eliminate the stream of payments reaped by the lords from their slaves. How would
masters be compensated for emancipating their slaves? As the elites holding power, masters were
likely to block any form of explicit abolition that failed to compensate them. Since slaves had no
assets or savings, they could not buy their freedom outright. A compensation scheme would
therefore have involved a form of a long-term contract. This time dimension creates (using
modern language) a double-sided commitment problem: what prevents the former slave from
escaping before paying compensation? What prevents the lord, who typically held a local
comparative advantage in violence, from reneging after payment has been made? The difficulty
2 Abbreviations are listed at the beginning of the references.
3 Gary Becker (1957) made a similar argument with respect to discrimination against minorities, the
American South in particular.
3
of solving these problems suggests that both sides would be wary of abolition that involved
compensation over time.
In modern, developed economies, the judicial system enforces contracts and the rule of
law more generally, affording credible solutions to double-sided commitment problems. The
legal system enforces long-term contracts.4 But most of the economies Smith discusses, such as
medieval Europe, lacked a functioning system of justice and the rule of law. This absence meant
that the methods available in modern, developed system to solve the commitment problem were
unavailable in the societies Smith discusses. Absent a means of compensation, slaveholders in
these societies rationally avoided abolition.
This paper proceeds as follows. Section 2 interprets Smith’s argument about inefficiency
in terms of the modern economics of slavery, as developed in the new economic history (Fogel
and Engerman 1974a, Goldin 1973, Wright 1978, 2006). In section 3, I discuss at greater length
the two arguments made by Smith as to why, despite inefficiency, slavery persists. A simple
game theoretic argument reveals the logic of the commitment problems hindering emancipation.
Section 4 analyzes Smith’s explanation for the abolition of slavery in Western Europe. In section
5, I draw on Smith’s arguments about political development in general to suggest that he had the
means to provide a third explanation for slavery’s persistence, another form of commitment
problem. Section 6 discusses Smith’s argument for why slavery is worse in democracies than
monarchies. My conclusions follow.
4 For this reason and because of the productive efficiencies of the gang labor system, slavery in the context
of the United States differed from the contexts studied by Smith.
4
2. The Smithian Economics of Slavery
Smith’s argument about slavery reported in the notes on his Lectures on Jurisprudence (LJ(A))
begins with an examination of incentives and the inefficiencies of slavery.
The slave or villain who cultivated the land cultivated it entirely for his master; whatever
it produced over and above his maintenance belonged to the landlord; he had therefore no
inducement to be at any great expense or trouble in manuring or tilling the land; if he
made it produce what was sufficient for his own maintenance this was all that he was
anxious about. The overseer perhaps by a hearty drubbing or other hard usage might
make him exert himself a little farther, so as to produce from the farm a small portion for
the landlord; but this would not be very great, and accordingly we see that a farm which
yielded 1/6 part of the produce to the master was reckoned to be tollerably well
cultivated. [LJ(A) 112-13:185-86]
Smith contrasts the lack of incentives for production under slavery with a free farmer who pays a
fixed rent to the landholder:
[A]s the free tenant pays a stated rent to the master, whatever he makes the farm produce
above that rent is intirely his own property, and the master can not exact as he could from
the ancient villains or slaves exact, any part they have saved above the rent, what they
had saved out of the part allowed for their maintenance. This gives them much greater
spirit and alacrity for their work; they will then be at expense to manure and improve
their land, and will soon bring it to that degree of cultivation as to be able to pay 1/3 part
to their masters and nevertheless have a much better as well as a more certain livelyhood
out of the remaining two thirds; and whatever they produce above that, which is supposed
to be about 1/3 of the produce, is altogether their own. Such a manner of cultivation is
therefore far I preferable to that by slaves, not only to the servants but even to the master.
[LJ(A) 113-14:186]
As the above passage attests, Smith concluded that free tenancy was vastly superior
economically to slavery. Using the quantities Smith’s provides above about production and
translating Smith’s arguments into modern economics, we have the following specific
conclusions about production and inefficiency.
Slavery forces slaves to live at the subsistence level and no more. As Smith asserts, the
production “over and above his maintenance belonged to the landlord.” Let S be the
subsistence level necessary to keep the slave alive. As Smith reports, “a farm which
5
yielded 1/6 part of the produce to the master was reckoned to be tollerably well
cultivated.” This implies that total produced from the land is , S to the slave for his
maintenance plus an additional 1/6 to the landlord.
Under free tenancy, total production is 3S. The free tenant pays 1/3 of the total, or S, to
the landlord. He keeps S for his maintenance, with the final S goes to the tenant for his
own use. This yields a net product of 2S.
The ratio of net production of free tenancy vs slavery is 12:1; i.e., 2S vs. S. The gross
product rises by a factor of more than 2.5: i.e., 3S vs.
.
From this discussion, we infer that free tenancy is a Pareto improvement over slavery.
Because free tenants face superior incentives, they produce so much more than slaves. According
to Smith’s figures, everyone should be better off under free tenancy. In the terms expressed by
Smith, the surplus captured by the landlord would increase by a factor of six (S vs. S/6) under
free tenancy; and the return to the tenant rises from 0 under slavery to S under free tenancy, a
huge increase above subsistence.5
3. The Puzzle: Why Did Slavery Persist?
The inference that freeing the slaves is Pareto superior to slavery raises an obvious puzzle: why
did slavery persist?
In the course of two long and complex paragraphs, Smith gives two very different
answers to this question. Smith’s initial answer is the “love of domination”:
Notwithstanding of these superior [of] advantages [of free labor] it is not likely that
slavery should be ever abolished... tho as I have here shewn their real interest would lead
5 N.B., Smith assumes that the production technology is identical under slavery and freedom. In contrast,
the gang-labor system in the Antebellum American South differed considerable from the tenant farming system that
arose following the Civil War and the abolition of slavery. The gang labor system made slavery more productive
than the alternatives for certain crops (Forgel and Engerman 1974). Therefore, Smith’s arguments do not apply to
slavery in the antebellum United States.
16
76S
76S
6
them to set free their slaves and cultivate their lands by free servants or tenents, yet the
love of domination and authority and the pleasure men take in having everything done by
their express orders, rather than to condescend to bargain and treat with those whom they
look upon as their inferiors and are inclined to use in a haughty way; this love of
domination and tyrannizing, I say, will make it impossible for the slaves in a free country
ever to recover their liberty.6 [LJ(A) 114:186]
Smith’s second answer involves a commitment problem involving whether masters could
reasonably expect to be compensated for their loss of slaves. In Smith’s words,
In all countries where slavery takes place[s] the greatest part of the riches of the subjects
consists in slaves. If he is possessed of a land estate the whole management of it is carried
on by the slaves; without them there can be nothing done; they work and till the ground,
and practise every thing else that is necessary to the cultivation of the land or the support
of their master. [LJ(A) III.115:187]
The power of the great lords consisted in their vassalls and their villains. The whole of
the land at this time was … cultivated by villains or slaves. [LJ(A) III.118:187-88]
Smith then raises the compensation problem:
To abolish slavery therefore would be to deprive the far greater part of the subjects, and
the nobles in particular, of the chief and most valuable part of their substance. This they
would never submit to, and a generall insurrection would ensue. For no single man ever
had or possibly could have power sufficient to enable him to strip his subjects in that
manner. If he set a slave at liberty this was robbing his master of the whole value of him.
This therefore could never take place. This institution therefore of slavery, which has
taken place in the beginning of every society, has hardly any possibility of being
abolished. [LJ(A) III.115-16:187]
Smith makes several arguments in this long passage detailing his second argument for the
stability of slavery. He explains that slaves are valuable assets. For even “tolerably” wealthy
landowners, “the greatest part of their wealth will consist” in slaves.7
6 Smith repeats the “love of dominance” thesis a few pages later, at [LJ(A) 129:192]; and again in his
discussion of this topic in WN: “The pride of man makes him love to domineer, and nothing mortifies him so much
as to be obliged to condescend to persuade his inferiors. Wherever the law allows it, and the nature of the work can
afford it, therefore, he will generally prefer the service of slaves to that of freemen.” [WN III.ii.10:388] 7 Smith also asserts that every subject – meaning the entire elite – had slaves. Moreover, a “man of a
considerable estate would have some thousands of slaves upon it, and the meaner sort in proportion, but allmost
7
Slaves as assets
I draw on both standard works on slavery in the “New Economic History” or NEH (see,
e.g., Fogel and Engreman 1974a,b; Golden 1973, Wright 1978, 2006) and on recent insights
from the theory of organizations/ institutions to interpret Smith’s logic in this paragraph.
Slaves generate a stream of produce for their masters. Using the quantities that Smith
provides, each slave produces an annual surplus of S/6 over and above that required for his own
maintenance (that is, S). The market value of a slave, V, is the discounted present value of the
stream of slave’s net produce.
To calculate V, let 0 ≤ δ ≤ 1 be the discount factor, a measure of the difference between
receiving a payoff of one today versus a payoff of one tomorrow. We then have:
V = ∑t=0,∞ δt[S/6] = (1)
S∕(6(1-δ)). (2)
Equality (1) says the following. The slave produces a net value for his master of S/6 in every
period. Because the value of receiving S/6 this year is higher than today’s value of the receipt of
S/6 next year, we discount next year’s receipt of S/6 by δ where 0 ≤ δ ≤ 1. This implies that
today’s present value of receiving S/6 next year is δ(S/6). Similarly, the receipt of S/6 in t years
from today is δt[S/6]. Hence, summing up the present value of the slave’s production for every
year, t, and discounting year t by δt, yields V, the discounted present value of the slave to its
owner. In economies where slaves are freely bought and sold, the market price of a slave is V.8
every one if the country be tollerably wealthy will have some slaves; and in them the greatest part of their wealth
will consist.” [LJ(A) III.115-16:187] 8 The NEH literature on slavery adds an additional level of complexity, allowing different types of slaves to
have different values (for example, children, prime field hands, women of child-bearing ages); I ignore this
complexity, though it could easily be added.
8
Algebraic manipulation allows us to simplify the value of V, as represented in the fraction
in expression (2). Notice that the larger is δ, the larger is V. For example, if δ = .95
(corresponding to a modest discount of next year of 5% and hence a long view of the future),
then V = 10S/3. If instead the discount factor is lower, say δ = .75 (corresponding to a large
discount of next year 25%, more highly discounting the future), then V = 2S/3.
These examples illustrate the main results. The discount factor significantly affects the
present value of a stream of income produced by an asset – in this case, a slave.9 In stable,
commercial environments without violence, the discount factor is likely to be quite high
(corresponding to low interest rates). But in less stable environments with significant violence,
discount factors are much lower (corresponding to higher interest rates). The calculations just
made illustrate the effect of this different in environments. In stable commercial environments
with high discount factors (e.g., δ = .95), the value of the asset is five times higher than in an
unstable environment with the threat of violence (e.g., δ = .75).
The compensation problem
Taking the productivity factors mentioned by Smith, freeing the slaves improves their
total output by a factor of 2.5 and their net product by a factor of 12. Let’s assume that in feudal
Europe landholders qua slaveholders represent the political elite and have a veto over any major
change, such as abolition.
9 That is,
V.
9
Two conditions must hold for the elites to favor abolition. First, as noted, the total surplus
must be sufficiently large to make the elite better off. Second, slaveholders must believe they
will, in fact, be better off under emancipation. In other words, for an individual master to favor
abolition, he must expect to receive at least V for each slave he frees. If the slaves are freed
without compensating masters, masters will be worse off, not better off. Because uncompensated
emancipation makes slaveholders worse off under abolition, they, as the elite, will block it. The
last quote from Smith explains this problem: “To abolish slavery therefore would be to deprive
the far greater part of the subjects, and the nobles in particular, of the chief and most valuable
part of their substance. This they would never submit to.” [LJ(A) III.116:187] Moreover, Smith
observes, “This institution therefore of slavery, which has taken place in the beginning of every
society … has hardly any possibility of being abolished.” [LJ(A) III.116:187]
Long-term compensation schemes
An alternative to immediate and uncompensated abolition of slavery involves a
compensation scheme whereby the freed slave works for a number of years, pays compensation
to the master, and is then free when the compensation totals an agreed upon price, P, where P ≥
V.10
Abolition in this scheme is a type of long-term contract.
Long-term compensation schemes in less-developed contexts, such as a feudal Europe,
are fraught with commitment problems. Each side would rationally worry that the other would
renege in some way. In the feudal environment with only a rudimentary government and no
10
Golden (1973) studies the economics of compensation schemes in the American context (see also Fogel
and Engerman 1974b). American law combines with a working judicial system to make the antebellum American
context considerably different from those Smith studied.
10
independent judiciary, contract enforcement could not rely on third parties enforcement of
contracts.
The government in the first stages of society is as I said11
very weak, and can not
interpose much in the affairs of individualls. Government is far advanced before the
legislative power can appoint judges at pleasure, as is now the case in Britain where the
king can appoint any one a judge he pleases who is a lawyer by profession, and for the
lower judiciall offices any one he pleases. This could not be done in an early society. The
people would not submit themselves in that manner. The government therefore would
find it necessary to take advantage of the superiority and authority of certain persons who
were respected in the country and put the judicial power into their hands. Jurisdictions
were in this manner established, and the same cause made it necessary to strengthen the
hands of all private masters of families. [LJ(A) III.116-17:187]
Smith makes two points in this paragraph. First, most of the societies with slavery are
insufficiently developed to have a judicial system capable of enforcing the rule of law.12
Second,
more subtly, the local lords not only ruled their domain, but served as its judge, including matters
in which the lord had a direct interest.
Consider in this context the commitment problems associated with the freed slave. First,
the freed slave’s commitment problem. If set free, former slaves would have little incentive to
stick around. Slaves have no love for their masters. If left to their own cognizance, they will be
tempted to leave without paying their debt, perhaps migrating to another region with better
opportunities. Setting the slaves free, granting them rights to make investments, accumulate
wealth and property, all increase a free man’s mobility. But if the former slaves leave, masters
will fail to recoup the value of their property.
Second, Lords also faced a commitment problem. In combination with their vassals, lords
commanded the lion’s share of the local violence potential. Nothing prevented the lord from